|
Early auditory responses to speech sounds in Parkinson’s disease: preliminary data
Scientific Reports volume 12, Article number: 1019 (2022) Cite this article
Abstract
Parkinson’s disease (PD), as a manifestation of basal ganglia dysfunction, is associated with a number of speech deficits, including reduced voice modulation and vocal output. Interestingly, previous work has shown that participants with PD show an increased feedback-driven motor response to unexpected fundamental frequency perturbations during speech production, and a heightened ability to detect differences in vocal pitch relative to control participants. Here, we explored one possible contributor to these enhanced responses. We recorded the frequency-following auditory brainstem response (FFR) to repetitions of the speech syllable [da] in PD and control participants. Participants with PD displayed a larger amplitude FFR related to the fundamental frequency of speech stimuli relative to the control group. The current preliminary results suggest the dysfunction of the basal ganglia in PD contributes to the early stage of auditory processing and may reflect one component of a broader sensorimotor processing impairment associated with the disease.
파킨슨병(PD)은
기저핵 기능 장애의 징후로서
음성 변조 및 음성 출력 감소를 비롯한
여러 가지 언어 장애와 관련이 있습니다.
voice modulation and vocal output
흥미롭게도
이전 연구에 따르면
PD를 가진 참가자는
음성 생성 중 예기치 않은 기본 주파수 교란에 대한 피드백 기반 운동 반응이 증가하고,
대조군 참가자에 비해 보컬 피치의 차이를 감지하는 능력이
향상되는 것으로 나타났습니다.
여기에서는
이러한 향상된 반응의 한 가지 가능한 원인을 살펴보았습니다.
우리는
PD와 대조군 참가자의 음성 음절 [da]의 반복에 대한
주파수 추종 청각 뇌간 반응(FFR)을 기록했습니다.
frequency-following auditory brainstem response
PD 참가자는
대조군에 비해 음성 자극의 기본 주파수와 관련된
더 큰 진폭의 FFR을 보였습니다.
현재의 예비 결과는
PD에서 기저핵의 기능 장애가
청각 처리의 초기 단계에 기여하며
이 질환과 관련된 광범위한 감각 운동 처리 장애의 한 구성 요소를 반영할 수 있음을 시사합니다.
Similar content being viewed by others
Atypical processing of tones and phonemes in Rett Syndrome as biomarkers of disease progression
Article Open access10 June 2020
Article Open access09 July 2024
Aberrant neurophysiological signaling associated with speech impairments in Parkinson’s disease
Article Open access14 April 2023
Introduction
Parkinson’s disease (PD), a manifestation of basal ganglia (BG) dysfunction, is associated with a number of speech production deficits in prosody, phonation and articulation, with phonation and laryngeal deficits the most prominent1,2,3,4,5. In addition to speech motor symptoms, auditory perceptual deficits ranging from self-monitoring to discrimination have been reported in PD5,6,7. Interestingly, it is during the monitoring of their own speech that PD participants often show the greatest differences from unimpaired speakers3,8,9. A common interpretation is that when individuals with PD are asked to produce speech with normal loudness (as judged by a speech-language pathologist), they perceive themselves as shouting or producing abnormally loud speech8. In addition, while listening at a given distance from a loudspeaker, individuals with PD estimated the loudness level to be significantly greater than that estimated by healthy control participants2.
기저핵(BG) 기능 장애의 징후인
파킨슨병(PD)은
운율, 발성 및 조음 prosody, phonation and articulation 에서
여러 가지 언어 생산 결함과 관련이 있으며,
발성 및 후두 결함이 가장 두드러지게 나타납니다1,2,3,4,5.
언어 운동 증상 외에도
자기 모니터링에서 변별에 이르는
from self-monitoring to discrimination
흥미롭게도,
PD 참가자들이 비장애 화자와 가장 큰 차이를 보이는 것은
일반적인 해석은
PD 환자에게
정상적인 음량으로 말을 산출하도록 요청할 때
자신이 소리를 지르거나
비정상적으로 큰 음량을 산출하는 것으로 인식 한다는 것입니다8.
또한
라우드스피커에서 일정한 거리를 두고 들을 때,
PD를 가진 사람들은 건강한 대조군 참가자들이 예상한 것보다
훨씬 더 큰 음량으로 평가했습니다2.
These perceptual/sensory differences and their relationship with the speech motor output deficits of PD have been examined in several recent studies. In response to alterations in auditory feedback during speech, PD participants display an interesting characteristic. When faced with feedback consistent with a misplaced tongue position for a specific vowel (change in first formant frequency of vowel), PD participants exhibit reduced compensation compared to age-matched control participants10, consistent with a weaker motor response. In contrast, when faced with feedback consistent with a change in fundamental frequency (fo, voice pitch), PD participants exhibit increased compensation to the feedback shift10,11,12,13. An increased response to voice pitch shifts has also been observed in participants with Alzheimer’s disease14,15 and cerebellar degeneration16. In a recent study, we demonstrated that the increased response to a change in voice pitch is accompanied also by increased sensitivity in detecting pitch alterations during auditory feedback monitoring in participants with PD4. Pitch feedback manipulations were presented under conditions of production and listening. In the production condition, participants’ vocal pitch was shifted, and participants judged whether their speech output had been manipulated in real-time; participants’ responses to pitch shift change were simultaneously recorded. During the listening task, participants judged whether paired tokens of their previously recorded speech samples were the same or different. Under the production condition, the ability of participants with PD to identify the pitch shift was greater than that of the controls, with a trend for better detection during the listening condition4. Interestingly, in a parallel experiment, detection accuracy of first formant shifts was reduced in individuals with PD only during the listening condition4.
이러한
지각/감각적 차이와 PD의 음성 운동 출력 결함과의 관계는
최근 여러 연구에서 조사되었습니다.
말하기 중 청각 피드백의 변화에 반응하여
PD 참가자들은 흥미로운 특징을 보입니다.
특정 모음에 대한 혀의 위치가
잘못 배치된 피드백(모음의 첫 번째 형태소 빈도 변화)에 직면했을 때,
PD 참가자는 연령이 일치하는 대조군 참가자10에 비해
운동 반응이 약해지면서
보상이 감소하는 것으로 나타났습니다.
반대로
기본 주파수(fo, 음성 피치)의 변화와 일치하는 피드백에 직면했을 때,
PD 참가자는 피드백 변화에 대한
보상이 증가합니다. 10, 11, 12, 13
소뇌 퇴행16이 있는 참가자에게서도
음성 피치 voice pitch 변화에 대한 반응이
증가하는 것이 관찰되었습니다.
최근 연구에서는
음성 피치 변화에 대한 반응이 증가하면
PD4 참가자의 청각 피드백 모니터링 중
피치 변화를 감지하는 민감도도 증가한다는 사실을 입증했습니다.
피치 피드백 조작은
제작 조건과 청취 조건에서 제시되었습니다.
생산 조건에서는 참가자의 음성 피치가 바뀌고 참가자는 실시간으로 음성 출력의 조작 여부를 판단했으며, 피치 변화에 대한 참가자의 반응이 동시에 기록되었습니다.
듣기 과제에서 참가자들은 이전에 녹음한 음성 샘플의 쌍을 이루는 토큰이 같은지 다른지 판단했습니다.
제작 조건에서
PD를 가진 참가자의 피치 시프트 식별 능력은
대조군보다 높았으며,
듣기 조건에서 더 잘 감지하는 경향을 보였습니다4.
흥미롭게도 병행 실험에서는
듣기 조건에서만 PD가 있는 참가자의 첫 번째 포먼트 이동 감지 정확도가 감소했습니다4.
At the neural level, electrophysiological recordings in response to voice pitch shifts yielded larger event related potentials in the inferior frontal gyrus, precentral gyrus, postcentral gyrus, and middle temporal gyrus in PD participants12. While the results suggest cortical involvement in the enhanced pitch shift response, there are also reports in the literature to suggest involvement at the level of the brainstem for auditory processing in PD participants17,18,19. The brainstem, including the cochlear nuclei, the superior olivary complex, and the inferior colliculus of the midbrain comprise the auditory pathway to auditory cortical areas, with processing at each level20. The main focus of the current investigation was to evaluate the brainstem involvement in the processing of complex speech sounds in relation to the speech motor deficits in PD.
신경 수준에서
음성 피치 변화에 대한 전기 생리학적 기록은
PD 참가자의 하전두이랑, 전전두이랑, 후전두이랑 및 중측두이랑에서
더 큰 사건 관련 전위를 산출했습니다12.
이 결과는
피질이 향상된 피치 이동 반응에 관여한다는 것을 시사하지만,
문헌에는 PD 참가자의 청각 처리를 위한
뇌간 수준에서의 관여를 시사하는 보고도 있습니다17,18,19.
달팽이관 핵, 중뇌의 상급 난소 복합체 및 하급 소체를 포함한 뇌간은
청각 피질 영역에 대한 청각 경로를 구성하며,
각 수준에서 처리됩니다20.
현재 조사의 주요 초점은
PD의 언어 운동 결손과 관련하여
복잡한 말소리 처리에서 뇌간이 관여하는 것을 평가하는 것이었습니다.
The auditory brainstem response (ABR) is an auditory evoked neural potential that provides a window into top–down and bottom–up processing of sensory information through the efferent and afferent projections of cortex, brainstem, and BG. Scalp-recorded auditory brainstem responses to complex sounds such as vowels include transient and sustained components that represent certain critical acoustic properties of speech stimuli21. Some of the acoustic properties of speech including pitch and formant frequencies are closely reflected within a component of the ABR (sustained response) known as the frequency-following response (FFR). The FFR reflects sustained electrical potentials that are precisely phase-locked to neuronal firing with an upper limit of about 1000 Hz in response to low to middle frequency periodic acoustic stimuli. Consequently, the FFR demonstrates a robust representation of time-varying fo and harmonics corresponding to the pitch and first formant frequency (F1) of the vowel (21,22 for a review refer to23). Measures that are derived from the FFR response, such as amplitude and latency, represent a mapping between the auditory stimulus and the neural activity, which may be modified due to changes associated with disease (e.g., autism spectrum disorders, and mild cognitive impairment24,25) or exposure to auditory stimulation (e.g., music, and bilingualism26,27). Specific regions in the BG and thalamic nuclei, including the subparafascicular thalamic nucleus, send dopaminergic projections to the main auditory midbrain nucleus and the inferior colliculus, consistent with a modulatory role of the BG in auditory processing28,29,30,31. Based on these considerations, it is suggested that a dysfunctional dopaminergic system may contribute to differences in the processing of pitch and formant information for speech and ultimately to differences in speech motor output.
청각 뇌간 반응 auditory brainstem response (ABR)은
청각 유발 신경 전위로,
피질, 뇌간 및 BG의 구심성 및 구심성 투영을 통해
감각 정보의 하향식 및 상향식 처리에 대한 창을 제공하는
청각 유발 신경 전위입니다.
모음과 같은 복잡한 소리에 대한
두피에 기록된 청각 뇌간 반응에는
음성 자극의 특정 중요한 음향 특성을 나타내는
일시적 및 지속적 구성 요소가 포함됩니다21.
피치 및 포먼트 주파수를 포함한
음성의 일부 음향 특성은
주파수 추종 반응(FFR)으로 알려진
ABR(지속 반응)의 구성 요소에 밀접하게 반영되어 있습니다.
FFR은 저주파에서 중주파의 주기적인 음향 자극에 대한 반응으로 약 1000Hz의 상한으로 뉴런 발화에 정확하게 위상이 고정된 지속 전위를 반영합니다. 결과적으로 FFR은 모음의 피치 및 첫 번째 포먼트 주파수(F1)에 해당하는 시간에 따라 변화하는 포먼트와 고조파를 강력하게 표현합니다(21,22 검토는 23 참조). 진폭 및 지연 시간과 같은 FFR 반응에서 파생된 측정값은 청각 자극과 신경 활동 간의 매핑을 나타내며, 이는 질병(예: 자폐 스펙트럼 장애 및 경도 인지 장애24,25) 또는 청각 자극 노출(예: 음악 및 이중 언어 사용26,27)과 관련된 변화로 인해 수정될 수 있습니다. 뇌하수체하 시상핵을 포함한 BG와 시상핵의 특정 영역은 청각 처리에서 BG의 조절 역할28,29,30,31과 일치하는 도파민성 투사를 주청각중뇌핵과 하대상회로 보냅니다.
이러한
고려 사항을 바탕으로 기능 장애가 있는 도파민 시스템은
음성에 대한 음높이 및 포먼트 정보 처리의 차이와
궁극적으로 음성 운동 출력의 차이에 기여할 수 있다고 제안합니다.
Here, we recorded FFR from the left mastoid during listening to a speech syllable. We compared patterns of phase-locking in individuals with PD and control participants on the assumption that it indexes—not solely, but to a large extent—subcortical interactions. On the basis of prior behavioral and perceptual results, we hypothesized that individuals with PD would show an increased FFR response only in the frequency range associated with fo compared to control participants, consistent with a brainstem contribution to their speech perceptual and production impairment.
여기에서는
음성 음절을 듣는 동안 왼쪽 유양돌기에서 FFR을
기록했습니다.
우리는 위상 고정이 전적으로, 아니 상당 부분 피질하 상호작용을 지표화한다는 가정하에 PD와 대조군 참가자의 위상 고정 패턴을 비교했습니다. 이전의 행동 및 지각 결과를 바탕으로, 우리는 PD를 가진 개인이 대조 참가자에 비해 fo와 관련된 주파수 범위에서만 FFR 반응이 증가하여 언어 지각 및 생산 장애에 대한 뇌간의 기여와 일치할 것이라는 가설을 세웠습니다.
Results
Electrophysiology
A multivariate analysis of variance (MANOVA) was used to compare root mean square (RMS) magnitudes (pre-stimulus and response) and spectral response amplitudes of the FFR (fo and F1) between groups (PD vs. control). A waveform depicting the characteristics of the input stimulus [da], along with averaged FFR waveforms for the PD and control groups, is displayed in Fig. 1a and b. The Figure also includes a bar graph comparing the magnitudes of response RMS amplitudes during the pre-stimulus and stimulus periods, and it is displayed in Fig. 1c. In the time domain, no significant difference was observed between groups for the amplitude of the pre-stimulus region [F(1,28) = 0.24, p = 0.626]. In contrast, individuals with PD showed greater RMS amplitude for the frequency-following response [F(1,28) = 9.04, p = 0.006, ηp2 = 0.244].
Figure 1
(a) Stimulus waveform. (b) Average waveform brainstem response to [da] in individuals with PD (red) and age- and gender-matched controls (black). The pre-stimulus period (− 10 to 0 ms) and the response period (0 to 60 ms) are shown. (c) Bar graph demonstrating between-group differences in RMS amplitude for the pre-stimulus and response periods. The error and shaded error bars represent the standard error. Note: *p < 0.01.
Figure 2 shows the averaged FFT (Fig. 2a) and bar graphs of the mean response amplitudes corresponding to fo and F1 (Fig. 2b) for the two groups. In the frequency domain, the amplitude of the frequency-following responses during the vowel portion of the stimulus (10–60 ms) for fo (80–120 Hz) and F1 (400–600 Hz) were assessed for group differences. Individuals with PD demonstrated larger amplitude FFR responses in the frequency range associated with fo compared to the control group [F(1,28) = 8.51, p = 0.007, ηp2 = 0.233; PD group (Mean (M) = 0.146, Standard Deviation (SD) = 0.047; Range = 0.04–0.23); Control group (M = 0.100, SD = 0.056; Range = 0.04–0.25)]. However, there was no group difference for the frequency range associated with F1 [F(1,28) = 0.002, p = 0.966; PD group (M = 0.014, SD = 0.009; Range = 0.01–0.04); Control group (M = 0.014, SD = 0.006; Range = 0.01–0.03)].
Figure 2
(a): Fast Fourier transforms calculated for the vowel part of the average response of individuals with PD (red) and age- and gender-matched control (black) group. (b) Bar graphs demonstrating between-group differences in response amplitude corresponding to fo (80–120 Hz) and F1 (400–600 Hz) representation of the speech stimuli. The error and shaded error bars represent the standard error. Note: *p < 0.01.
Correlation analysis
To further investigate the relationship between the neurophysiological response to the auditory speech stimuli and the severity of behavioural speech disorders (dysarthria) of PD, we correlated the FFR amplitudes corresponding to fo and F1 with the Movement Disorder Society Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) and perceptual dysarthria ratings for each participant with PD. No significant correlation was observed between the amplitudes corresponding to fo or F1 and the UPDRS (fo and MDS-UPDRS: r(13) = 0.088, p = 0.754; F1 and MDS-UPDRS: r(13) = 0.153, p = 0.587) or the perceptual dysarthria scores (fo and dysarthria score: r(13) = − 0.368, p = 0.177; F1 and dysarthria score: r(13) = − 0.428, p = 0.111).
Discussion
Our preliminary results are consistent with an enhanced encoding of vocal pitch, evidenced by increased FFR amplitude in the frequency range of 80–120 Hz, in individuals with PD compared to age- and gender-matched control participants. In contrast, we did not find evidence of enhanced encoding within the range of the first formant (400–600 Hz). The increased amplitude of the FFR response is consistent with a selective modulation of the fundamental frequency of the speech stimuli at the level of the brainstem. The brainstem components of the FFR are modulated by the BG through inhibitory and disinhibitory projections to auditory relay areas including the inferior colliculus (IC30), with the IC one of the largest generators of the FFR response32,33,34.
Increased encoding of vocal pitch has been observed at the cortical level evidenced by larger auditory evoked potentials (P2 responses) in PD compared to non-PD participants12. Since the brainstem FFR can be modulated from the cortex35, the specific source of the enhancement (brainstem or cortex) can't be determined from the present results. However, it appears that the increased FFR response in PD participants reflects a reduction in the inhibitory (or tuning) function of the auditory brainstem pathway which enhances the frequency response within the vocal pitch range. In our previous study, PD participants were better able to detect and compensate for pitch shifts compared to the non-PD participants4. The current results suggest that the difference in FFR amplitude in the pitch range enhances the salience of the acoustic signal enabling better detection of differences in pitch during listening. In addition, the enhanced feedback signal interacts with the motor output during voice production to produce a stronger response to pitch shifts.
BG damage and its influence on the auditory signal on vocal output during production may be accompanied by an opposite change to the sensitivity of the somatosensory system. Voice production generates auditory as well as somatosensory feedback resulting from the laryngeal vibration, and feedback from both sensory systems contribute to the perception of pitch and loudness36,37. For individuals with no history of neurodegenerative disease, vocal fold mucous anesthetization yields a greater compensatory response to auditory feedback alterations in pitch compared to the pitch response without anesthesia38, suggesting a trade-off in sensitivity or gain between the two sources of sensory input (somatosensory vs. auditory). Similarly, masking of laryngeal somatosensory feedback by applying low-pass filtered stochastic vibrations to the neck enhances the Lombard response (increased loudness)37. In individuals with PD, using air-puff stimulation, it was found that thresholds for detecting laryngeal somatosensory input are increased, consistent with reduced sensitivity of mechanoreceptors in the laryngeal area39. As a result, BG damage appears to influence both sensory modalities (auditory and somatosensory) in opposing ways resulting in an increase in sensitivity to pitch-related auditory feedback and a decrease in sensitivity to pitch-related somatosensory feedback.
One consequence of the opposing changes in feedback from the two sensory systems is an imbalance in their contributions to speech production. The addition of the pitch shift would combine with normal feedback to increase the perception of pitch and loudness, while the change in threshold for somatosensory input would result in smaller movements due to the reduced reafferent input. Interestingly, one therapy for remediation of the hypokinetic features in the speech of individuals with PD involves vocal exercises focused on the production of loud speech (Lee Silverman Voice Treatment—LSVT9,40,41,42,43). The approach generates greater somatosensory feedback from the articulatory and laryngeal systems compared to normal speech. One of the effects of the LSVT is smaller vocal pitch compensations following treatment suggesting a down regulation of the enhanced auditory response44. Hence, it is reasonable to suggest that by increasing loudness, the LSVT treatment may act to reset the balance between the feedback systems providing the conditions for more normal speech production and perception.
In our previous work, we also found a reduction in compensation to manipulations of F1 auditory feedback10, and a reduction in F1 error detection during listening in our PD participants4. In the present study there was no evidence of an enhancement or a reduction in the FFR at frequencies in the range of the first formant. The lack of difference in FFR amplitude in the first formant frequency range suggests that the motor and sensory deficits related to F1 may occur at a higher level than the brainstem. Because formant information is associated with phonetic processing, it more likely involves primary and association levels of the processing stream, which rely on integration of both formant and harmonic information to extract relevant linguistic information. As a result, the influence of BG dysfunction on speech has differential effects related to the processing levels for different components of the speech production (and presumably) the perceptual process.
This study has some shortcomings that should be mentioned. First, most of the individuals with PD in the current study showed mild to moderate severity. It would be of interest to study a greater range of severity in future studies to have a more complete picture of the effect of PD on speech auditory brainstem processing. It is also important to consider other concomitant factors along with PD on the FFR responses, such as amount of musical training or multilingualism (in the case of tonal languages as the second language), as there are reports of the effects of both on FFR responses26,27. In order to recruit as many participants as possible, we did not screen for the amount of musical training or multilingualism in the current study, these factors merit attention in future investigations. In addition, based on recent work on different speech subtypes of PD45,46, it would be beneficial to assess the relationship of the sensory deficits and different speech subtypes. In the current work, we did not classify different speech subtypes of individuals with PD due to the relatively small sample size. In the future, one can, for example, investigate whether pitch and loudness sensory differences are more frequent in the prosodic subtype of speech deficits in PD as opposed to the phonatory-prosodic or articulatory-prosodic subtypes.
In conclusion, in these preliminary data, we found increased frequency-following neural responses related to fo during the perception of speech in individuals with PD compared to age- and gender-matched control participants. These findings provide a neural basis for the sensory processing deficits of vocal pitch and loudness at the brainstem level in this population10,11,13. Impaired modulation of sensory information at the BG may be one possible factor in the manifestations of speech deficits in individuals with PD.
토론
우리의 예비 결과는 연령과 성별이 일치하는 대조군 참가자에 비해 PD를 가진 개인에서 80-120Hz의 주파수 범위에서 FFR 진폭이 증가한 것으로 입증된 보컬 피치의 향상된 인코딩과 일치합니다. 반면, 첫 번째 포먼트(400-600Hz) 범위 내에서는 인코딩이 향상되었다는 증거를 찾지 못했습니다. FFR 반응의 진폭 증가는 뇌간 수준에서 음성 자극의 기본 주파수를 선택적으로 변조하는 것과 일치합니다. FFR의 뇌간 구성 요소는 하부 콜리큘러스(IC30)를 포함한 청각 중계 영역에 대한 억제 및 비억제 투영을 통해 BG에 의해 변조되며, IC는 FFR 반응의 가장 큰 생성자 중 하나입니다32,33,34.
피질 수준에서 보컬 피치의 인코딩 증가가 관찰되었는데, 이는 비PD 참가자에 비해 PD에서 더 큰 청각 유발 전위(P2 반응)를 통해 입증되었습니다12. 뇌간 FFR은 피질에서 변조될 수 있기 때문에35, 현재 결과로는 강화의 특정 소스(뇌간 또는 피질)를 확인할 수 없습니다. 그러나 PD 참가자의 FFR 반응 증가는 음성 피치 범위 내에서 주파수 반응을 향상시키는 청각 뇌간 경로의 억제(또는 튜닝) 기능의 감소를 반영하는 것으로 보입니다. 이전 연구에서 PD 참가자는 비PD 참가자에 비해 음정 변화를 더 잘 감지하고 보정할 수 있었습니다4. 이번 연구 결과는 피치 범위에서 FFR 진폭의 차이가 음향 신호의 명료성을 향상시켜 청취 중 피치 차이를 더 잘 감지할 수 있음을 시사합니다. 또한 향상된 피드백 신호는 음성 생성 중에 모터 출력과 상호 작용하여 피치 변화에 대한 더 강력한 반응을 생성합니다.
BG 손상과 이것이 음성 생산 중 청각 신호에 미치는 영향은 체성감각 시스템의 감도와 반대되는 변화를 동반할 수 있습니다. 음성 생산은 후두 진동으로 인한 청각뿐만 아니라 체성감각 피드백을 생성하며, 두 감각 시스템의 피드백은 음조와 음량을 인식하는 데 기여합니다36,37. 신경 퇴행성 질환의 병력이 없는 사람의 경우 성대 점막 마취를 하면 마취를 하지 않은 음정 반응에 비해 청각 피드백 변화에 대한 보상 반응이 더 크게 나타나며38, 이는 두 감각 입력원(체성 감각 대 청각) 간의 민감도 또는 이득의 균형을 시사합니다. 마찬가지로, 저역 통과 필터링된 확률적 진동을 목에 적용하여 후두 체성 감각 피드백을 마스킹하면 롬바드 반응(음량 증가)이향상됩니다37. PD 환자의 경우, 에어 퍼프 자극을 사용하면 후두 부위의 기계 수용체의 감도 감소와 일치하여 후두 체성 감각 입력을 감지하는 임계값이 증가하는 것으로 밝혀졌습니다39. 결과적으로 BG 손상은 두 감각 양식(청각 및 체성감각)에 반대되는 방식으로 영향을 미쳐 음조 관련 청각 피드백에 대한 민감도는 증가하고 음조 관련 체성감각 피드백에 대한 민감도는 감소하는 것으로 보입니다.
두 감각 시스템의 피드백이 상반된 방식으로 변화하는 결과 중 하나는 음성 생산에 대한 두 감각 시스템의 기여도가 불균형해지는 것입니다. 음조 변화는 정상적인 피드백과 결합하여 음조와 음량에 대한 인식을 증가시키는 반면, 체성 감각 입력의 역치 변화는 재귀 입력 감소로 인해 더 작은 움직임을 초래할 수 있습니다. 흥미롭게도, PD 환자의 말에서 저운동성 특징을 개선하기 위한 한 가지 치료법에는 큰 소리 내기에 초점을 맞춘 발성 운동이 포함됩니다(리 실버만 음성 치료-LSVT9,40,41,42,43). 이 접근 방식은 정상 발성에 비해 조음 및 후두 시스템에서 더 큰 체성감각 피드백을 생성합니다. LSVT의 효과 중 하나는 향상된 청각 반응의 하향 조절을 시사하는 치료 후 보컬 피치 보상이 작아진다는 것입니다44. 따라서 음량을 증가시킴으로써 LSVT 치료가 보다 정상적인 음성 생산과 지각을 위한 조건을 제공하는 피드백 시스템 간의 균형을 재설정하는 역할을 할 수 있다고 제안하는 것이 합리적입니다.
이전 연구에서는 F1청각 피드백 조작에 대한 보상 감소10와 PD 참가자의 청취 중 F1오류 감지가 감소하는 것을 발견했습니다4. 본 연구에서는 첫 번째 포먼트 범위의 주파수에서 FFR의 향상 또는 감소에 대한 증거가 없었습니다. 첫 번째 포먼트 주파수 범위에서 FFR 진폭의 차이가 없다는 것은 F1과관련된 운동 및 감각 결손이 뇌간보다 더 높은 수준에서 발생할 수 있음을 시사합니다. 포먼트 정보는 음성 처리와 관련이 있기 때문에 관련 언어 정보를 추출하기 위해 포먼트 및 고조파 정보의 통합에 의존하는 처리 스트림의 기본 및 연관 수준을 포함할 가능성이 더 높습니다. 결과적으로 BG 기능 장애가 음성에 미치는 영향은 음성 생산의 여러 구성 요소(그리고 아마도 지각 과정)에 대한 처리 수준과 관련된 차별적인 영향을 미칩니다.
이 연구에서는 언급해야 할 몇 가지 단점이 있습니다. 첫째, 이번 연구에 참여한 대부분의 PD 환자들은 경증에서 중등도의 중증도를 보였습니다. 향후 연구에서는 더 넓은 범위의 중증도를 연구하여 PD가 언어 청각 뇌간 처리에 미치는 영향에 대한 보다 완전한 그림을 그리는 것이 흥미로울 것입니다. 또한 음악 훈련의 양이나 다국어 사용(제2언어인 음조 언어의 경우)이 FFR 반응에 미치는 영향에 대한 보고가 있으므로 PD와 함께 다른 수반되는 요인도 고려하는 것이 중요합니다26,27. 가능한 한 많은 참가자를 모집하기 위해 이번 연구에서는 음악 훈련의 양이나 다국어 구사 능력에 대한 선별을 하지 않았으므로 향후 조사에서 이러한 요소에 주목할 필요가 있습니다. 또한 PD45,46의 다양한 언어 하위 유형에 대한 최근 연구를 바탕으로 감각 결핍과 다양한 언어 하위 유형의 관계를 평가하는 것이 도움이 될 것입니다. 현재 연구에서는 표본 규모가 상대적으로 작아 PD 환자의 다양한 언어 하위 유형을 분류하지 않았습니다. 향후에는 예를 들어, PD에서 음성 결함의 운율 하위 유형에서 음높이 및 음량 감각 차이가 음성-운율 또는 조음-운율 하위 유형과 달리 더 자주 발생하는지 여부를 조사할 수 있습니다.
결론적으로, 이러한 예비 데이터에서 우리는 연령과 성별이 일치하는 대조군 참가자에 비해 PD 환자의 언어 지각 시 주파수 추종 신경 반응이 증가하는 것을 발견했습니다.
이러한 발견은
이 집단에서 뇌간 수준에서 보컬 피치 및 음량의 감각 처리 결함에 대한
BG에서
감각 정보의 변조 장애는
PD 환자의 언어 결함 증상의 한 가지 가능한 요인일 수 있습니다.
Materials and methods
Ethics statement
This study was approved by the McGill Faculty of Medicine Institutional Review Board, in accordance with principles expressed in the Declaration of Helsinki. Informed written consent was obtained from participants prior to their involvement in the research project.
Participants
Fifteen patients with Parkinson’s disease (6 female, 9 male; mean age: 65.87 years) and fifteen age- and gender-matched control participants (6 female, 9 male; mean age: 63.13 years) were recruited for this study (same group as our previous study4,10). The severity of PD motor symptoms, assessed using the Movement Disorder Society Unified Parkinson’s Disease Rating Scale (MDS-UPDRS; Part III Motor Examination47), ranged from mild (a score of 13) to moderate (a score of 48; mean [M] ± standard deviation [SD] score, 24.79 ± 9.19). Cognitive functioning was assessed using the Montreal Cognitive Assessment (MoCA48) and was in the normal range for all individuals with PD (scores > 26). All patients were taking L-dopa in addition to other medication, including dopaminergic and/or anticholinergic drugs. Participants were tested off medication for 12 h. Two participants reported a history of speech therapy focused on increasing speech loudness and intelligibility.
Each participant read aloud the Rainbow Passage (a standard speech perceptual passage assessment) in order to carry out a perceptual analysis of dysarthric speech characteristics. A licensed Speech-Language Pathologist rated the speech of participants on 43 perceptual dimensions that span the speech subsystems, including phonatory and articulatory subsystems, using a 7-point scale1. Overall, the severity of participants with PD was rated as moderate (2 participants), mild-to-moderate (4 participants), mild (6 participants), and within normal limits (3 participants). Inter-rater agreement was tested between the first rater and a second listener using intraclass correlation (ICC) in order to assess consistency in the ratings of speech perceptual characteristics in individuals with PD. The resulting ICC was in the excellent range, ICC = 0.9049, indicating that the raters had a high degree of agreement. We used the perceptual and MDS-UPDRS scores to evaluate any relationship between severity of speech and motor symptoms and the magnitude of the FFR responses.
All participants underwent an audiometric screening and were found to have binaural pure tone hearing thresholds of 40 dB HL or less at 250, 500, 1000, 2000 and 4000 Hz. None of the participants used hearing aids. All participants were native speakers of North-American English. Participants in the control group were healthy with no history of neurological condition.
Stimulus and recording
A 40 ms (ms) speech syllable, [da], was synthesized at a 20 kHz sampling rate using a Klatt synthesizer50. After a 5 ms stop burst, voicing remained constant with a fundamental frequency of 100 Hz, and the first formant frequency of 500 Hz. The [da] stimulus was chosen because it combines transient ([d], the first 10 ms with a 5 ms voice onset time) and periodic ([a]) segments51), two acoustic features that have been extensively studied in speech ABR52. For each participant, the [da] stimulus was presented 12,000 times with a 50 ms interstimulus-interval. Stimuli were presented in alternating stimulus polarities (i.e., compression and decompression of air molecules of periodic sound waves: positive and negative) to both ears at 80 dB SPL through electromagnetically-shielded insert earphones (Etymotic ER-2) to reduce stimulus and noise artifacts, using a TDT stimulus presentation system (Tucker-Davis Technologies, TDT Inc., Alachua, FL). A vertical montage of four electrodes (left mastoid active, two on the forehead as grounds, and a hairline reference) was used, with all impedances kept under 5 kΩ. Continuous responses were recorded (20 kHz sampling frequency) with ActiABR200 software (BioSemi, Amsterdam, Netherlands). During the recording session (lasting approximately 18 min), participants sat in a comfortable chair in a sound attenuated room.
Data processing
Electrophysiological responses were band-pass filtered offline in EEGLAB53 between 70 and 2000 Hz to maximize signal-to-noise ratio and detection of peaks within the phase-locking limits of the brainstem. The root mean square (RMS) amplitude of the neural responses was used to quantify the overall magnitude of response and pre-stimulus activity. RMS amplitudes were computed for the pre-stimulus period (− 10 to 0 ms) and the response period (0 to 60 ms). A fast Fourier transform (FFT) was performed on a signal window between 10 and 60 ms, corresponding to the voiced portion of the stimulus (Brainstem toolbox52). Mean amplitude across frequency ranges corresponding to fo (80 to 120 Hz) and F1 (400 to 600 Hz) were calculated for each trial and then averaged across participants for each group. Responses were then averaged over a − 10 to 60 ms window, with stimulus onset occurring at time zero. Any trial with an amplitude greater than 40 µV was considered an artifact and rejected before averaging. We included 11,400 trials of response averages (5700 trials in each polarity) in the analysis after artifact rejection across participants. There were no differences in the number of rejected trials between the two groups [t(28) = 0.971, p = 2.048]. Responses (0 to 60 ms) were then amplitude baseline-corrected in the pre-stimulus period (− 10 to 0 ms). In a final step, responses from the two stimulus polarities were averaged to minimize the influence of cochlear microphonic and stimulus artifact on the measured response54.
Statistical analysis
Statistical analyses were performed using multivariate analysis of variance (MANOVA) with root mean square (RMS) magnitudes (pre-stimulus and response) and spectral responses (fo and F1) as within-subject factors and group (PD vs. control) as between-subject factor. Factor and simple effect sizes were quantified using ηp2 to assess any statistically significant effects, defined as small (0.2–0.3), medium (0.5), and large (> 0.855). Greenhouse–Geisser corrections for unequal variances were applied when necessary. In addition, separate Pearson correlation analyses were performed between FFR amplitudes of fo or F1 and perceptual or MDS-UPDRS clinical scores in individuals with PD. This resulted in four correlation analyses. A Bonferroni-adjusted α rate of 0.012 was used.
Data availability
Anonymized datasets recorded and analyzed during this study are available from the corresponding author within the limits of participants’ consent.
References
|